Other Wise

The future is not our’s you see

I previously simplified the implemented definition of a future down to:

  1. Buy goods now to be delivered at time X
  2. Sell goods now to be delivered at time Q

That was a mistake. A necessary element of the core problem of Modern Mogul is that fact that the players should (nearly) never be playing the game as it exists in present time. Instead they should be playing the game as they hope it will be some number of turns in the future. In the above simplified definition the capital is both committed and spent now and only satisfaction is delayed. The more interesting and useful form has the commitment occur now and the capital expenditure when the future matures. In this way the players will not only have to juggle the vagaries of the future market, but also the vagaries of their future liquidity as a result of market activity between then and now. Much better.

Recycled whirlpools

A thought model; just a skein of partial ideas to consider hanging a game from:

External markets phase

  • A graph, nodes connected by edges. Some nodes are identified as (potential) start nodes.
  • Nodes bear a die indicating the current market demand at that node. cf The future market die (central collumn) in Lokomotive Werks
    • The die or node may be coloured to indicate a specific transport type (passenger, freight etc)
  • Players may claim edges between start nodes and other nodes, or between nodes already connected and other nodes (connected or unconnected). In this manner the network grows and becomes incestuous
    • Possibly an auction of the nominated edge?
    • cf Dutch Intercity’s blind-bid edge within a very small graph
    • The Riding Series method of rotation auctions until at least one player has spent all their cash and all players have auctioned at least once may be interesting
    • Ordering by cash holdings, or inverse, may also be interesting. cf The Riding Series and Lokomotive Werks for value considerations
    • Possibly there are limitations on contiguous networks, costs for non-contiguous networks etc.
    • Should claimed edges be coloured by player and if so, what is the value?
  • By claiming a connection a player has acquired commits them to supply transport in that volume in N-turns (N = 3?)
    • The auction for the edge sets the price for that product-type ($/volume) and this is tracked/recorded
    • Future auctions produce new values, also tracked
    • Minimum bids may be set as a function of projected values from this history
  • As additional edges are claimed new markets enter the game with their own demands for volume
    • Represented by a new “line” of dice for that node, possibly in a different colour or a higher value of a current colour
    • Thus a thriving market for one sort of transport (eg freight) prompts a market for passengers etc etc and so forth (yay theme)
  • In a later phase players may sell the futures among each other
    • Again an auction, modifies the tracked price?

Production phase

  • Effectively Lokomotive Werks redressed
    • D4s? — lower variance
    • Number of dice in future orders column equal to total of all dice-values on nodes connected by claimed edges
  • Building new factories unchanged.
  • Building capacity unchanged
  • Players don’t claim dice, they auction them (and receive the money personally)
    • What is auctioned is a commitment to supply Q product of that type in N turned (N=3?)

Future fulfilment

  • The current turn’s contracts are resolved
  • Each buy must be matched with sells of the same type etc.
    • Open negotiation?
  • Unmatched buys are satisfied by the hidden market which operates at a cost of X% of a function of the tracked market prices and those funds are covered by the players
  • Unmatched sells force the market price down and penalise the holding player by n% of the then market value?
    • Simply losing the future may be enough penalty
  • What is resolution order of what buys and sells are matched and in what order?
  • Monies are paid from the bank to players holding satisfied buy futures (they ship the goods?)

Market maintenance phase

  • Each dice line on a node grows or shrinks by one die Lokomotive Werks-style
    • Possibly this is a global function across all nodes, a local function/property of the node, or a mix of both
    • Possibly a function of the properties/types of lines in neighbouring nodes (reflective/communicative/memetic markets)?
    • The dice are rolled, establishing total orders within the period (may be less or more than total futures)
  • Each player in rotation claims either auctions a die at a currently connected node (assumes uncoloured edges?), or auctions a new/additional edge
    • A new edge may connect a new node, in which case they automatically get that die/contract, or if between already-connected nodes, may affect market maturation patterns in those nodes (see above)
    • Not clear what auction value an edge between connected nodes has?
  • Limit to total growth of dice lines on nodes?

A fractured thought model to be sure, plangent and struck through with gaping flaws.

Returning to the future

After some odd thoughts about how the network growth potentials of ‘Ohana Proa could be mapped as a pseudo-futures market, i’ve been thinking about representing futures in games quite a bit lately and I’m having a really hard time of it.

The simplest way to present a future seems to be as a pair of contracts:

  1. The agreement to sell at time X for price Y
  2. The agreement to buy at time Q for price R

The two are of course reciprocals. There are more complex forms of futures, but they are just that, more complex. The basic form would have PlayerA selling a contract to PlayerB, PlayerB giving PlayerA money in return for the guaranteed market (to buy or sell). The complexity is that PlayerB has to have the money now to buy the contract, and depending on the type of the contract, will also either have to have the trade goods or the money to satisfy the contract when it comes due. The primary justification for futures is to guarantee both the pricing and the existence of the marketplace into the future against the vagaries of market dynamics. Thus a wheat producer can not only set his effective price for next year’s tonnage now, but can also guarantee that there is in fact a market to sell his wheat next year. A wheat buyer can do likewise, protecting themselves against price spikes due to bad weather and a reduced crop (for instance). But you know all this — I repeat myself, re-tracing the pattern in hope that repetition reveals something new and useful.

The challenge is to (literally?) transcribe this into a game. There must be risks sufficiently large that the inefficiencies of futures are worth it. The range of extent of the risks must also be knowable (supply may be high or low, prices may be high or low, demand may be high or low). Additionally, if a futures market is to be represented, the actual trade of futures and the tracking of futures prices as a discrete market entities, then fulfilment must be sufficiently postponed that the value of the future has the opportunity for multiple significant changes and can be reasonably traded multiple times.

This suggests a future length of between 3 and 5 rounds1. The future would be sold and there would then be 3-5 rounds of market evolutions and potential trading of a given future before it matures and is fulfilled. A good rule-of-thumb is that a game has 7-10 rounds, with each player making at least one strategically significant evolution per round. 7-10 rounds gives sufficient time for player investments to mature, carrying player commitment and value, and for game arc to develop. Given that futures must have a maturation time of 3-5 rounds and assuming that all the maturation periods will stack as densely as possible, that means a game duration of 10-15 rounds.

This of course assumes that the players start the game with positions which require the sort of risk abeyance that futures provide: another problem to resolve. But, so far so good. The remaining problems are to establish the risks and the market variance controls which can express reasonably in 10-15 rounds.

The future definition could also be simplified to :

  1. Buy goods now to be delivered at time X
  2. Sell goods now to be delivered at time Q

In essence the transaction occurs now except the fulfilment, the actual movement of goods, happens later. The pattern is similar to the mail order market model (given guaranteed delivery times) except that fulfilment is delayed by more than the postal system2. This is almost the same kettle of fish except that all the capital is required up front for the purchaser to abey the risks.

If the simplified model is combined with a loans/temporary-liquidity system 3 it becomes remarkably similar to the more classical and complex split contract/fulfilment/payment futures model.

Now for the risk system, and this is where I’m running dry.

The system needs a market with the following properties:

  1. High potential price variance
  2. Increasingly difficult to predict the longer the prediction (early-, mid- or late-game), not exponentially, but probably worse than linearly
  3. Reasonably predictable market volume demands

That last is a doozie. If I simulate supply and demand then either the supply and demand is highly unpredictable, or there is another large random factor in addition to supply and demand which affects market prices. One appealing possibility is using Lokomotive Werks’ novel dice-based market system, as it does provide a highly dynamic market with high variance supply and demand relationships4.

Positing Modern Mogul as an extrapolation of Lokomotive Werks is appealing as Lokomotive Werks is a fine game (review). Lokomotive Werks internally simulates the demand side through a combination of the number of dice rolled (which is a function of competition), the values that are rolled on those dice and the individual player’s turn order but leaves the supply side for players to construct. Somehow this needs to be split so that players can occupy both the supply and demand sides of the equation.

Demands are supplies for the next stemp in the consumption pipeline. What if there were the equivalent of two Lokomotive Werk’s tracks, one of factories for production (much like the current game) and the other with second stage demand? Thus, keeping with the train/transport themeing, one system would generate erratically growing sequences of transport demands (passenger, livestock, freight, etc), perhaps using a very similar system to Lokomotive Werks, which the players would then attempt to fulfil by representing back to the train production market as demands in some value distribution. Thus on the one hand players would attempt to gauge their position against the variations in transport demand, and by reflection from that, against the variance in production vehicle demand.

Much as I like perfect and certain information games, the use of dice (as a stand-in for any random system) seems suitable. here. The requirement is to create a high-risk difficult to predict system. A random system obviously does that and has the advantage that the range and distribution can be relatively precisely tailored. Building the system from player-interactions is likely possible but seems a difficult butterfly effect challenge and runs the risk of being fragile/crackable.

Hurm. Back-to-back Lokomotive Werk’s-like systems with a full futures market (time-based contracts to buy-and sell, plus dynamic price tracking). This needs musing on.

Harrumph. I started writing this post indending to explain how impossible it was to reasonably represent a futures market in a game and I had a stack of good solid (swiss-cheese) reasons in-hand. However, as hoped, as I assembled and articulated the reasons they built something else.

Footnotes
  1. A round being defined as a well bounded set of player actions and decisions sufficient to cause a significant strategic position change for each player
  2. I’m purposefully going to ignore delivery location.
  3. cf Container or Wealth of Nations
  4. This is rather cute, as Lokomotive Werks effectively implies a futures market for the trains the players build while never actually implementing any portion of that market!

Feeling up the future

The core elements of a future’s market are:

  1. The sale of the agreement to purchase a set quantity at a set price at a set time in the future
  2. The sale of the agreement to sell a set quantity at a set price at a set time in the future

That’s it: The purchased promise to transact volume at price on date. Future’s markets rely on two things:

  1. A reasonable certainty on market prices in the future
  2. Unpredictable and yet reasonable risks which could significantly affect the prices at that time

Wealth of Nations delivers (a poor) something of a futures market in the way that commodity prices vary with supply and demand (which is clever and works well), but does not allow direct speculative exploitation of that market. I’m specifically interested in a direct speculatively exploitative futures market in a game…

It seems difficult but not impossible to implement direct representation of a futures market in a game if the game is going to play quickly (less than 2 hours). If the players are to build the systems which establishes the reasonable certainty then I suspect it may be impossible as too many iterations are required to build the player investments to deliver that certainty — with fewer iterations the sunk costs of chaos are too low.

Perhaps if the game provided the backdrop of both the base market and the base forces which drive the general curces of future market behaviour? Thus the game would effectively guarantee that the basic curves of value and time were as so, but the players could affect those curves or deliberately cause events to throw them off? Hurm. Curious. That puts the players more in the role of market manipulators, attempting to buy one side of the speculation prospects up cheap before forcing the market in that direction. Hurm. Yeah.

So how would the market represent? How could players see the value curves they would be speculating on, let alone the player manipulation incentive grids for exploitation? That’s a problem.

Imagine a linear track of value. Each commodity has five markers on the track:

  1. Historical market high (plus a date/age?)
  2. Historical market low (plus a date/age?)
  3. Current price
  4. Price at the end of last turn/epoch
  5. Price at the end of the turn/epoch before last

That would seem to sum most of the really basic data needed for market behaviours: trend, range, and velocity. (Later update: Of course it misses volume, which is also key)

I wonder?

This is such a curiously ungainly area for a game.

Introducing the very modern major mogul

This is a temporarily named place holder with apologies to Gilbert & Sullivan. It is all in early gestation and may yet be stillborn.

Core problem: Iterative exploitation of leverage opportunities across a three-layered market: Production/sales, transport, futures.

Core concepts: factories, multi-typed transport (road, rail, sea/canal, air etc), 18XX technology/obsolescence, player-grown undirected graph of transport-coloured edges, speculative futures market, demand-based pricing market, hedge operations

In short this is a melange of notions that muddled through the shampoo while in the shower.